<h2>Must know</h2>
<ul>
<li><strong>WhatsApp swaps C++ for Rust for media security</strong>: WhatsApp has replaced 160,000 lines of C++ code with 90,000 lines of Rust to create Kaleidoscope, a new media consistency library. This library validates media files and detects malicious patterns, protecting billions of users from malware hidden in files. According to Facebook Engineering, this is the largest global deployment of Rust code for a consumer application, demonstrating Rust’s production-readiness and improving performance and memory usage across Android, iOS, web, and wearables. <a href="https://app.daily.dev/posts/2YIPmtHiF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Ingress-nginx project archived, time to migrate to Cilium</strong>: The ingress-nginx project is being archived in early 2026, meaning no more active maintenance or security updates. Users need to migrate to alternatives like Cilium Ingress for a direct replacement or Cilium’s Gateway API for advanced features. CNCF reports that both options leverage Cilium’s eBPF-powered datapath for better performance and observability, with migration tools available to ease the transition. <a href="https://app.daily.dev/posts/PptBbhnAu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft Office zero-day vulnerability actively exploited</strong>: Microsoft issued an emergency patch for CVE-2026-21509, a zero-day vulnerability in Office with a CVSS score of 7.8. This flaw allows attackers to bypass security features for legacy COM and OLE components using malicious Office files. While newer Office versions received immediate updates, Office 2016 and 2019 users need manual registry workarounds. CISA has mandated patching for US federal agencies by February 16, highlighting the urgency of this actively exploited threat. <a href="https://app.daily.dev/feed-by-ids?id=qC937DXHe&amp;id=qC937DXHe" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Google adds visual reasoning to Gemini 3 Flash</strong>: Google has introduced Agentic Vision capabilities to its Gemini 3 Flash model, allowing it to interactively analyze images, execute Python code for annotations, and generate visualizations from complex data like high-density tables. <a href="https://app.daily.dev/posts/KIgS0LnaZ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rust Foundation releases 2025 annual report and strategic plan</strong>: The Rust Foundation released its 2025 Annual Report and 2026-2028 strategic plan, prioritizing stable infrastructure, sustainable maintainer support, responsible adoption growth, and community engagement, having raised $5.1 million in financial support for the Rust Project in 2025. <a href="https://app.daily.dev/posts/DRpplNBlF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Xfce developers fund Wayland compositor project</strong>: Xfce developers are initiating the development of Xfwl4, a new Wayland compositor for the Xfce desktop environment, funded by community donations and built in Rust for memory safety and modern graphics support. <a href="https://app.daily.dev/feed-by-ids?id=f6P6uC82l&amp;id=96AM3XRbN&amp;id=96AM3XRbN&amp;id=96AM3XRbN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel project adopts formal succession plan</strong>: The Linux kernel project has formalized a succession plan for leadership after Linus Torvalds’ eventual departure, ensuring continuity and addressing concerns about an aging maintainer community. <a href="https://app.daily.dev/posts/vAKZlS9Qz" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google Cloud launches N4A instances with Axion ARM CPUs</strong>: Google Cloud has released N4A compute instances powered by custom Axion ARM processors, offering up to 64 vCPUs and 512GB RAM, targeting web servers, microservices, and containerized applications with improved performance and cost-effectiveness. <a href="https://app.daily.dev/feed-by-ids?id=JvBjVyZ5U&amp;id=JvBjVyZ5U" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New terminal UI tool Overitall launches</strong>: Overitall (oit) is a new terminal UI tool for process and log management, combining interactive log filtering, trace detection, and VS Code integration, with explicit support for AI assistants like Claude. <a href="https://app.daily.dev/feed-by-ids?id=DkhXBDdU9&amp;id=DkhXBDdU9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Vercel CLI adds API command for direct API access</strong>: Vercel CLI version 50.5.1 introduces a new <code>api</code> command, allowing authenticated HTTP requests to Vercel APIs directly from the terminal, supporting both manual users and AI agents like Claude Code. <a href="https://app.daily.dev/feed-by-ids?id=AvpZcZC0U&amp;id=AvpZcZC0U" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-generated code: the fast fashion of software</strong>: Apparently, ‘vibe coding’—generating tons of AI code without human review—is leading to a ‘false sense of productivity’ and ‘overly complex code that is difficult to maintain.’ <a href="http://Fast.ai" target="_blank" rel="noopener nofollow">Fast.ai</a>, the source of this profound insight, claims developers overestimate their productivity by 40%. So, while AI is supposed to make us faster, it’s actually just creating more garbage for us to clean up. Who knew? <a href="https://app.daily.dev/feed-by-ids?id=fPZ532s1X&amp;id=fPZ532s1X" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cursor AI browser project: built from scratch, or just a pile of spaghetti?</strong>: Cursor announced a ‘major AI software engineering breakthrough’—an AI-built web browser with ‘over 3 million lines of code’ and a ‘custom Rust-based rendering engine.’ Developers, however, quickly found the code to be ‘spaghetti,’ unable to compile, and heavily reliant on existing projects. So much for ‘from scratch’ and ‘fully working.’ Sounds like another AI hype cycle delivering more promises than actual product. <a href="https://app.daily.dev/feed-by-ids?id=gcgtwbk5H&amp;id=gcgtwbk5H&amp;id=gcgtwbk5H" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ChatGPT’s new favorite encyclopedia: AI-generated Grokipedia</strong>: The Guardian found ChatGPT’s GPT-5.2 model citing Elon Musk’s AI-generated Grokipedia as a source, raising concerns about amplifying bias and misinformation. OpenAI claims its models filter low-credibility sources, but apparently, an AI-generated encyclopedia passes the test. This recursive AI sourcing could lead to ‘model collapse,’ where AI models just keep feeding on each other’s hallucinations. What could possibly go wrong? <a href="https://app.daily.dev/posts/fWRmbZwbT" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
